International Dental Journal (2010) 60, 7-30

Oral health risks of tobacco use and effects of cessation

Saman Warnakulasuriya1, Thomas Dietrich2, Michael M. Bornstein3, Elias Casals Peidró4, Philip M. Preshaw5, Clemens Walter6, Jan L Wennström7, Jan Bergström8

King’s College London and WHO Collaborating Centre for Oral Cancer and Precancer, UK, 2Department of Oral Surgery, University of Birmingham, UK; 3Department of Oral Surgery and Stomatology, University of Bern, Switzerland; 4Department of Preventive & Community Dentistry, University of Barcelona, Spain; 5Department of Periodontology, School of Dental Sciences, Newcastle University, UK; 6Department of Periodontology, Endodontology and Cariology, University of Basel, Switzerland; 7Department of Periodontology, Institute of Odontology, The Sahlgrenska Academy at University of Gothenburg, Sweden; 8Karolinska Institute, Sweden. 1

The purpose of this paper is to review the epidemiologic evidence for the effects of tobacco use and tobacco use cessation on a variety of oral diseases and conditions. Exposures considered include cigarette and bidi smoking, pipe and cigar smoking, and smokeless tobacco use. Oral diseases and disorders considered include oral cancer and precancer, periodontal disease, caries and tooth loss, gingival recession and other benign mucosal disorders as well as implant failure. Particular attention is given to the impact of tobacco use cessation on oral health outcomes. We conclude that robust epidemiologic evidence exists for adverse oral health effects of tobacco smoking and other types of tobacco use. In addition, there is compelling evidence to support significant benefits of tobacco use cessation with regard to various oral health outcomes. Substantial oral health benefits can be expected from abstention and successful smoking cessation in a variety of populations across all ages. Key words: Smoking, smokeless tobacco, oral cancer, pre cancer, periodontal disease, tooth loss, implants, dental caries, smoking cessation

Tobacco is used in a variety of ways, mostly as smoked, but many populations use smokeless tobacco, which comes in two main forms; snuff (finely ground or cut tobacco leaves that can be dry or moist, loose or portion packed in sachets) and chewing tobacco (loose leaf, in pouches of tobacco leaves, plug or twist form). This review examines the oral health risks of both smoked and smokeless tobacco. This literature review aims to present published evidence regarding our current understanding of the epidemiology, aetiology and pathogenesis of tobacco use-related disorders. In addition we also review significant improvements in oral health following cessation. The focus of the review is on the adverse effects of tobacco on several oral disorders including oral cancer, other oral mucosal disorders, periodontal disease and tooth loss, and how tobacco affects clinical management © 2010 FDI/World Dental Press 0020-6539/10/01007-24

such as implantology, and to discuss the oral health benefits of tobacco cessation. Smoking Oral cancer

Among sites that have been considered to be at highest relative risk for cancer due to smoking is the lung. Following lung cancer the highest relative risks are observed for the larynx and oral cavity1. The risk of oral cancer has increased in recent decades in many countries in the world2. In those countries in which epidemiological studies have been conducted, it is clear that oral cancer risk is high among smokers. A recent meta-analysis reported 12 studies that estimated oral cancer risk in the USA, Uruguay, Italy, Sweden, India, China, Taiwan doi:10.1922/IDJ_2532Warnakulasuriya24

8

and Korea3. The reported pooled cancer risk estimate was 3.43 times higher in smokers compared with nonsmokers (95% CI 2.37, 4.94) (Figure 1). The results for risks associated with tobacco smoking were generally consistent across countries entered into the metaanalysis except in the study conducted among females in Sweden4. In a study reported from Northern Italy, the single factor with the highest attributable risk was smoking, which accounted for 81-87% of oral cancers in males and for 42-47% in females5. It is evident that oral cancer risk is related to both intensity and duration of tobacco smoking. The differential risk between nonsmokers and heavy smokers, and the steady progression of risk with increasing amount smoked both provide sufficient evidence for tobacco as a major risk factor for oral cancer. Furthermore most studies show an inverse relation with age when starting to smoke. Among young people in southern England, a significant risk among males (alcohol adjusted OR: 19.5, 95%CI 1.3,286.8) was associated with starting to smoke under the age of 16

years6. These risks are also increased synergistically with alcohol consumption. However, among never drinkers, cigarette smoking was associated with an increased risk of 2.13 (95% CI 1.52,2.98) confirming an independent association with tobacco use7. This had also been demonstrated in an earlier study among 19 cases and 213 controls who described themselves as non-drinkers; the ORs were 3.8 (95%CI 0.2, 58.2) and 12.9 (95%CI 2.3,106.3) for smokers of 20% teeth with PD 4+mm

>10% teeth with CAL 5+mm

2.4

Kibayashi33

219

3+ sites with 2+mm increase 3+ sites with 2+mm increase

2.3

Linden47

1362

AL 6+mm

15% sites with AL 6+mm

Nicolau36

251

low AL

high AL

Thomson41

810

1+ site(s) with AL 4+mm

1+ site(s) with AL 4+mm

Wang44

1590

Do32

3161

Do32

500*

Vered43

7056

40

1.4 1.3

1.7

3.9

2.8

1.2

2.3

3.3

3.5

2.6

4.8

2.2

2.6

5

7.1

1.5 2+ sites with AL 4+mm or PD 5+mm 2+ site(s) with AL 6+mm and PD 5+ mm CPI = 3 or 4

* Subset with severe periodontitis

International Dental Journal (2010) Vol. 60/No.1

CPI = 3 or 4

3.3 1.6

1.2

1.6

1.6

1.8

4.7

3.3

11

It can further be seen from Table 1 that the sample size of the studies varies considerably from 40 participants in a clinical study of therapeutic outcomes to 7,056 individuals in a screening study of army personnel. On average, the number of participants per study was 1,567, similar to that reported in the previous review27. The authors of these studies assessed the periodontal condition by means of a variety of clinical and/or radiographic measures such as pocket depth, attachment level (loss), and bone height (loss). A few studies used the number of retained (or lost) teeth as the endpoint of choice. Most studies made use of more than one clinical measure for assessment of the periodontal condition and three studies used a compound measure, the Community Periodontal Index (CPI), which relies on pocket depth. The most commonly used clinical measures or endpoints are pocket depth and attachment level (loss). The outcome with regard to the influence of smoking, however, is independent of endpoint, which is in good agreement with the observations of the review referred27. The fact that observations based on different criteria or measures used for disease assessment arrive at the same conclusion, again, increases the generalisabilty of the findings. Pocket depth is a frequently used measure for the assessment of the periodontal clinical condition. Although common and accepted, pocket depth may be a problematic measure when studying periodontal disease in relation to smoking. Pocket depth measurements are influenced by the quantity and quality of the inflammatory tissue at the apical base of the pocket. The inflammatory tissue at the base of the pocket is commonly tighter and less extensive in the smoker compared with a corresponding non-smoker, which results in less ‘probe penetration’ in the smoker46. This may be a problem in studies comparing pocket depth measures between smokers and non-smokers with a risk for underestimating the depth among smokers relative to non-smokers. In addition, gingival recession is commonly more extensive in smokers resulting in a reduced pocket depth when compared with the corresponding situation in a non-smoker. Both elements of the pocket depth measurement process could lead to underestimation of the severity of the clinical condition in a smoker. Two of the more recent articles in this review have presented data that are inconclusive with respect to the influence of smoking on periodontal disease30,42. Although the objective of the present review was not primarily to scrutinise in detail the quality of single studies, a closer look at these two might be relevant. One of the studies30 concerns elderly persons and suffers severely from the fact that tooth loss was high in this population (23% were edentulous). Furthermore, only 36% out of 473 persons were found with complete dental records, and only 10% of these persons were smokers (17 persons). It seems unrealistic even to try to get reliable information from such data. Moreover, studies based

on elderly persons, in general, are problematic regarding the exploration of smoking effects, since premature mortality at older ages affects smokers more than nonsmokers. Elderly smokers eligible to participate in such a study will most likely not be representative of the wider population, and the effect of smoking, therefore, could be underestimated. The other study is a 15-year longitudinal followup of an Indonesian population cohort42. The main problem with this study is the immense (50%) loss to follow-up. In addition, a problem particularly related to smoking was encountered: whereas nearly all men were smokers, women were not. Under such extreme circumstances the effect of smoking will most likely be confounded by gender. An analysis restricted to males would have been instructive in such a situation. The shortcomings of these two studies illustrate some of the difficulties that may be met when estimating the influence of smoking on periodontal health. Risk estimation

The relative risk for periodontal disease associated with smoking was assessed in 11 of the 18 new studies: seven cross-sectional30,32,36,37,40,43,44,47 and four cohort33,38,41,42 studies. A summary is presented in Table 2. The magnitude of the relative risk estimates varied from 1.4 to 5.0 in the different studies. A considerable variation of the magnitude of the relative risk estimates was also observed in a previous review48. There may be several explanations for such variation, the main ones probably being the definitions of periodontal disease as well as smoking. In all the studies that have estimated the smoking associated relative risk at more than one level of disease severity, the relative risk increases as the severity level increases (Table 2)30,32,33,36-38,40,41,43,44,47. Thus, as the criteria become stricter, indicating more severe disease, the lower the prevalence and the greater the proportion of smokers49. Hence, the smoking associated risk increases as the prevalence of periodontal disease decreases. Furthermore, data from studies that have estimated the smoking associated relative risk at more than one level of smoking exposure indicate that the relative risk for periodontal disease in light smokers, on average, is less than that for heavy smokers (Table 2). This suggests an exposure gradient or a ‘dose-response’ relationship. Similarly, most of the studies that have included former smokers32,37,38,40,41,47 suggest that former smokers have a better periodontal condition or less relative risk than comparable smokers who continue to smoke, which points in the same direction. Both circumstances, increasing risk with increasing disease severity and increasing risk with increasing exposure, favour a causal effect by smoking. In two of the cohort studies, estimates of the incident risk were presented38,41. Surprisingly little information is available about periodontal disease incidence, i.e., the occurrence of new cases of the disease among Warnakulasuriya et al.: Oral health risks of tobacco use and effects of cessation

12

previously unaffected persons. This lack of knowledge has now been remedied to some extent from these two studies. According to these observations the 4-year38 and the 6-year41 cumulative incident risk for periodontal disease was 1.7-fold and 5.2-fold, respectively, higher in smokers. These two studies show that smokers, in addition to carrying a larger periodontal disease burden, also develop the disease earlier than non-smokers. This novel information is important and again favours a causal effect of smoking on periodontal disease. Thus, it is clear that more recently published articles on the topic of smoking and periodontal disease truly confirm earlier evidence and expand our understanding that smoking exerts a strong untoward effect on periodontal health and is a major risk factor for periodontal disease. Mechanisms of periodontal disease causation

The periodontal tissues are continuously exposed to nicotine and its metabolites due to deposition of nicotine on the root surface50 and cotinine levels (a metabolite of nicotine) are elevated in saliva and gingival crevicular fluid51. The current knowledge on periodontal disease causation by smoking has been summarised in excellent reviews52-55 in which several aetiological effects of tobacco smoke are discussed. However, tobacco related periodontal pathogenesis is not well understood. Some potential mechanisms are briefly presented here. The importance of oral micro-organisms for the development of periodontal inflammation was proven about 40 years ago56. Cigarette smoking is likely to affect the composition of the oral microflora due to a decrease in oxygen tension in periodontal pockets and may lead to a selection of anaerobic bacteria57. However, the literature is inconclusive concerning a specific smoking associated microbial profile58,59. The oral bacteria are organised within biofilms. In addition, the oral microflora consists of more than 700 different types of bacteria and many of these are not classified and cannot be cultured60. Periodontal diseases are likely to be associated with different microbial profiles, rather than to be associated with distinct pathogenic bacteria. The pathogenic subgingival biofilm has both direct and indirect effects on the periodontal tissues. In fact, damage to the periodontium results even without influence of bacterial invasion into the corresponding periodontal tissues because of the immune response of the host to bacterial stimulation61. Tobacco smoking affects the humoral mediated and the cell mediated immunity of the host and this may increase susceptibility to periodontal disease53,62-66. However, available data are conflicting and precise mechanisms have yet to be confirmed. The periodontal tissues are very well vascularised67. Typical signs of an inflammation, such as changes in gingival colour, swelling of the marginal as well as papillary gingiva, an increase of gingival crevicular fluid International Dental Journal (2010) Vol. 60/No.1

flow as well as bleeding on gentle periodontal probing (BOP) are caused by alterations of the vascular system. In smokers the clinical signs of inflammation and BOP are suppressed68,69. The literature is inconclusive concerning a clear smoking associated pathohistological correlate70-73 but clearly, periodontal inflammatory responses are altered as a result of smoking. There is evidence for an impact of smoking on bone metabolism such as an increased secretion of the bone resorbing factors PGE2 and IL-1β74 or a decreased intestinal uptake of calcium75, and these factors may also increase susceptibility to periodontal disease in smokers. Periodontal disease is influenced by genetic factors76. There is some evidence that tobacco smoking may affect the genetically determined susceptibility for periodontal diseases77-79, though again, precise mechanisms remain to be elucidated. Smoking-induced periodontal destruction results from a wide range of effects of tobacco on the different functions of cells, tissues and organs. Some of these effects are opposed to each other due to the effects of different tobacco constituents. However, when summarising the characteristics of tobacco-induced alterations on periodontal tissues and humoral immunity, it is very likely that tobacco smoking disrupts the physiological turnover of tooth-supporting structures with the net effect being periodontal tissue breakdown53,64,80. Smoking and tooth loss

Cigarette smoking has been shown to be associated with fewer remaining teeth in a plethora of cross-sectional studies in various populations35,81-88. Furthermore, increased rates of tooth loss among smokers have been observed in a number of longitudinal studies84,89-94. Among 789 men followed for up to 35 years in the VA Dental Longitudinal Study (VA DLS), an ongoing closed panel longitudinal study of men in the Greater Boston area (USA), rates of tooth loss among current cigarette smokers were approximately twice those of never smokers (adjusted HR: 2.1, 95% CI 1.5, 3.1)94. In another study conducted in Boston, the relative risk of tooth loss in current smokers compared to never smokers was 3.4 (95% CI 2.1, 5.7) among 248 dentate women followed over 2 to 7 years84. Okamoto et al. followed a total of 1,332 adult Japanese men over four years and found that current smokers were approximately twice as likely to experience tooth loss as never smokers38. In a study of 1,031 Swedish women followed over 12 years, women who smoked at baseline lost a mean number of 3.5 teeth compared to a mean of 2.1 teeth lost among never smokers89. Results of the largest prospective cohort study on cigarette smoking and tooth loss were recently reported by Dietrich et al.90. In this study, a total of 43,112 US male health professionals, including 26,284 dentists, were observed from 1986 to 2002, and data on smoking behaviour and tooth loss were updated

13

Dose-dependant association of current smoking with risk of tooth loss (Reference: never smokers)

Hazard-Ratio (95% CI)

4

3

2

1

0

1-4

5 – 14

15 - 24

25 - 34

35 - 44

45+

number of cigarettes per day © Thomas Dietrich

Figure 2. Hazard ratios and 95% confidence intervals (CI) for the dose-dependent association between current cigarette smoking and risk of tooth loss in the Health Professionals Followup Study, United States 1986-2002. (used with permission). HR adjusted for age, race, BMI, physical activity, diabetes, profession, routine medical examination, alcohol consumption, caloric intake, multivitamin use and vitamin C supplement use.

Table 3 Systematic reviews on implant failures in smokers versus non-smokers. Author Year

Type of study

Groups

Outcome

Level of analysis Results

Hinode et al. 2006 97

Systematic review (up to August 2005);

Smokers vs non-smokers

Removed implant

Implant level

Smokers vs non-smokers

Implant survival Implant level

OR 2.17 (95% CI 1.67-2.83) Significant effect of smoking only in the maxilla

12 case-control and 7 cohort studies (retro-/prospective) Mean follow-up: NR Klokkevold & Han 2007 98

Strietzel et al. 2007 99

Systematic review (up to May 2005);

Smokers 89.7% (95%CI 87.092.4)

14 case-control studies (retro-/ prospective)

Non-smokers 93.3% (95%CI 91.0-95.6)

Mean follow-up: 4.7 years

Pooled estimate of difference in implant survival 2.68% (95%CI 1.1-4.26)

Systematic review (up to December 2005); 29 case-control and cohort studies (retro-/prospective) Mean follow-up: NR

Smokers vs non-smokers

Removed implant, bone loss >50%, mobility, persistent pain or periimplantitis

Implant level (19 OR 2.17 (95% CI 1.67-2.83) studies) OR 2.64 (95% CI 1.70-4.09) Subject level (10 No significant difference in OR studies) for early (1 year) failures

NR: not reported OR: odds ratio

Warnakulasuriya et al.: Oral health risks of tobacco use and effects of cessation

14 Table 4 Publications after the time period covered by the systematic reviews described in Table 3. Groups Nunmber of patients/ implants Years of follow-up

Author Year

Type of study

Moy et al. 2005104

Retrospective 1140/4680 0.5-21 years

Smokers vs non-smokers

Outcome

Level of analysis

Subject level Removed implant, bone loss >50%, mobility, persistent pain or peri-implantitis

Risk of failure / % failed

Comments

Risk ratio: 1.56 (95%CI 1.03-2.36)

Majority of failures within the first year

Mundt et al. 2006105 Retrospective 157/663 Mean 7.3 years

Removed Smokers, exsmokers, never- implant smokers

Implant level

Wagenberg & Froum 2006107

Retrospective 891/1925 1-16 years; Mean 6 years

Smokers, nonsmokers

Removed implant

Implant level

Smokers 5.6% Non-smokers 3.7% (p=0.34)

No significant difference between rough- and smooth-surface implants.

Ellegard et al. 2006 100

Prospective cohort

Smokers, nonsmokers

Removed implant

Implant level

Hazard rate 2.2 (95%CI 0.8-6.1)

Sinus membrane lift

Kinsel & Liss 2007 102

Retrospective Smokers 12/95 Smoker (≥20 Non-smokers cig/day), nonsmokers 95/249 2-10 years

Removed implant

Subject level

Smokers 5/12; 42% Non-smokers 5/31; 16% (p=0.11) Smokers 7.4% Non-smokers 3.6%

Immediate loading protocol

Smokers OR 4.6 (95% CI 1.1-19) Ex-smokers OR 0.42 (95%CI 0.1-12.1)

No association between smoking habits and implant loss.

Implant level History of progressive periimplant bone loss (bone loss ≥3 threads after year 1)

OR 2.2 (95%CI 1.53.3)

No data on implant loss in smokers versus nonsmokers

Primary implant Implant level stability

OR 1.36 (95%CI 0.87/2.12)

5.1% of the implants removed before 2nd stage surgery. No data on implant loss in smokers versus nonsmokers

68/262 0-12.6 years; Mean 5.4 years

Subject level

Implant level

Roos-Jansåker et al. Cross-sec2006 106 tional

218/999 9-14 years

Peri-implantitis Subject level Smokers, exsmokers, never- (bone loss ≥3 threads after smokers year 1 and pus/ bleeding on probing)

Fransson et al. 2008101

82/ 439 Mean 9.4 years

Smokers, nonsmokers

Cross-sectional

Mesa et al. 2008 103 Retrospective 316/1084 Smoker (≥20 Up to 2nd stage cig/day), nonsurgery (early smokers failures)

OR: odds ratio

International Dental Journal (2010) Vol. 60/No.1

Smokers 15%, Ex-smoker 9.6% Never-smokers 3.6% (sign. difference between all groups) Sign association between duration of smoking and implant failure

15

biannually. The study demonstrated a dose-dependent association between current smoking and incidence of tooth loss, with heavy smokers having three times higher rates of tooth loss than never smokers (Figure 2). Cigar and pipe smoking are also likely to be related to tooth loss risk, but very few studies have investigated this hypothesis. A cross-sectional study of 705 individuals enrolled in the Baltimore Longitudinal Study of Aging reported a higher number of missing teeth among smokers of pipes and cigars compared to nonsmokers95. Two longitudinal studies in predominantly white US males have analysed pipe and cigar smoking as determinants of tooth loss risk90,96. In the VA DLS, both cigar and pipe smoking were independently associated with increased tooth loss risk in a longitudinal study of 690 men96. Among US health professionals, current pipe or cigar smoking was independently associated with tooth loss incidence compared to never or former smokers of pipes or cigars (HR: 1.20, 95% CI: 1.11, 1.30)90. Because the reference category in this study included former smokers of pipe and cigars some of whom may have quit relatively recently, current pipe and cigar smoking may actually elevate the risk of tooth loss by more than 20%; however, the effect is likely to be small compared to the effect of cigarette smoking. Implant failure

The influence of smoking on the short and long-term outcomes of implant therapy has been addressed in three recent systematic literature reviews (Table 3)97-99. The reviews cover the literature up to the year 2005. A literature search performed in PubMed covering the time period 2006 until June 2008 identified a further eight articles addressing the issue of outcomes of implant therapy in smokers (Table 4)100-107. Three of these articles did not have implant removal (or implant survival) as primary outcome variable, but described prevalence of peri-implantitis106, progressive bone loss101 or primary implant stability103 The systematic review by Hinode et al.97 used ‘removed implant’ (without specifying reasons for removal) as the outcome. The analysis included 19 studies and revealed that implants in smokers had a significantly higher risk to be lost (OR: 2.17, 95%CI 1.67,2.83), and also that the increased risk was significant only for implants placed in the maxilla. Klokkevold and Han98 included 14 studies in their systematic review and found the implant survival rate to be lower in smokers than non-smokers; pooled estimate of risk 2.68 (95%CI 1.1,4.26) over a mean follow-up of 4.7 years. While the two systematic reviews referred to above described data at the implant level, Strietzel et al.99 performed analyses at both the subject and implant levels. In this review, a broader definition of implant failure was used and included removed implant, bone loss >50%, mobility, persistent pain or peri-implantitis. Based on informa-

tion from 10 studies providing data at the subject level, smokers showed an overall OR of 2.64 (95% CI 1.70, 4.09) to experience implant failure. Five of the 10 included studies (50%) reported a statistically significantly increased risk for implant failure in smokers. The corresponding OR at the implant level (18 studies included) was 2.17 (95% CI 1.67, 2.83). No significant difference in OR for early (50%, mobility, persistent pain or peri-implantitis), demonstrated a higher risk of implant failure among smokers (HR: 1.56, 95% CI 1.03,2.36) and with the majority of the failures occurring within the first year of function104. A study evaluating primary implant stability103, however, revealed no significant difference between heavy smokers (≥20 cig./day) and non-smokers. Two cross-sectional studies focused on peri-implantitis9 or history of progressive peri-implant bone loss101 as the outcome variable and for each showed a significantly higher prevalence in smokers than in non-smokers, OR 4.6 (95% CI 1.1,19) and 2.2 (95% CI 1.5, 3.3), respectively. The evaluation of the current literature indicates that implant failure is more common among smokers than non-smokers, although the scientific evidence must be considered limited. A general problem in the analysis of the literature regarding the influence of smoking on the short- and long-term outcome of implant therapy is the insufficient quality of the studies with regard to design and/or description. Since smoking is a subject related factor, analysis of the effect of smoking habits on the outcome of implant therapy should be performed at the subject level. However, a majority of the studies include evaluation only at the implant level. Further, in most studies the data have been collected retrospectively and usually analysed using bivariate statistical methods without considering potential confounding factors (e.g. severity of periodontal disease, standard of hygiene, maxillary/mandibular jaw, characteristics of implant surface). Hence, in order to be able to properly define the effect of smoking on the prognosis of implant therapy there is an obvious need for adequately designed prospective studies. Although the proven beneficial effects of smoking cessation in general are known, the effect of smoking cessation intervention on implant survival rates should be evaluated.

Warnakulasuriya et al.: Oral health risks of tobacco use and effects of cessation

16

Dental caries

Few cross-sectional studies have explored the relationship between smoking and dental caries, and of these, many have not considered the confounding factors in sufficient detail. A study of military personnel from Illinois (mean age=25.9 years), found that smokers had a significantly higher number of untreated decayed teeth, missing surfaces and total DMFS than non-smokers (DMFS of 19.38 for non-smokers compared to 24.59 for current smokers)108. No information was obtained regarding diet or oral hygiene. In a sample of 1,156 elderly patients, Jette et al.109 also found a relationship between smoking (current) and dental decay, ORs (compared with all smoking categories combined) being 1.47 for smokers (95% CI 1.00,2.17), 1.02 for ex-smokers (95% CI 0.78,1.34) and 0.67 for never smokers (95% CI 0.50,0.89). In a longitudinal study which followed patients undergoing periodontal therapy by a specialist for a period of 12 years, Ravald et al.110 found that smokers developed significantly (p20 years) ex-smokers tended to be lower for mouth

(RR:1.61) than for laryngeal cancer (RR:3.63)190. These data suggest that for ex-smokers, the relative risk for oral cancer declines with time since stopping smoking, to almost reach the levels observed for non smokers. In a Swedish study, patients with oral leukoplakia who were smokers were asked to give up their smoking. It was found that leukoplakias present in persons with smoking habits were reversible when the smoking habit was reduced or given up195. In a 10-year follow up study in India the cessation of tobacco use led to a substantial fall in the incidence of oral leukoplakia200. Effects of smoking cessation on the periodontium

There are few articles in the periodontal literature which specifically address the impact of smoking cessation on the periodontium, presumably because of the inherent difficulties in motivating patients to quit smoking. Most investigations of the impact of smoking cessation on the periodontium have been either cross-sectional comparisons of periodontal status in current smokers, former smokers and never smokers, or have been prospective cohort studies. To date there has only been one intervention study to assess the effect of smoking cessation on periodontal treatment outcomes (described in more detail below)194. The periodontal literature that has focused on smoking cessation can therefore be divided depending on the type of study conducted. Cross-sectional and prospective cohort studies can be split into two categories: the effect of smoking cessation on periodontal status and the effect of smoking cessation on the outcome of periodontal therapy. Intervention studies of the impact of smoking cessation on the periodontium constitute their own category and include studies of the biological effects of quitting smoking. For more background, the reader is referred to excellent reviews by Tonetti201, Ramseier202, Heasman203 and Johnson80. Warnakulasuriya et al.: Oral health risks of tobacco use and effects of cessation

20

The effect of smoking cessation on periodontal status

In broad terms, studies have indicated that periodontal status in former smokers is usually intermediate to that of current smokers and never smokers. Current smokers usually have significantly worse periodontal status (greater probing depths, attachment loss, and alveolar bone loss) than either former smokers or never smokers. Several studies have indicated that the periodontal status of former smokers, while intermediate to that of current smokers or never smokers, is usually closer to that of never smokers than current smokers. Such findings provide indirect evidence of the periodontal benefits of quitting smoking and permit the conclusion that not only should smoking status always be assessed as a key parameter to indicate the periodontal disease risk for an individual patient, but smoking cessation counselling should form an integral part of periodontal therapy201. The principal findings of studies that have investigated the effect of smoking cessation on periodontal status are shown in Table 632,41,93,109,191-193,204-211. The effect of smoking cessation on periodontal treatment outcomes

A small number of studies have investigated the efficacy of periodontal treatment in former smokers as compared to current smokers or never smokers, and these also provide indirect evidence of the benefit of smoking cessation on periodontal treatment outcomes. Taken collectively these studies confirm that treatment outcomes in former smokers are generally similar (if sometimes slightly inferior) to those that can be expected in never smokers, but are usually better than those that can be expected in current smokers. The principle findings of these studies are shown in Table 7212-217. Intervention studies

A small number of intervention studies have been conducted that have investigated the effect of smoking cessation on the periodontium. These studies have indicated that smoking has a negative impact on the gingival microvasculature and that these changes are reversible on quitting smoking. To date, only one interventional study to assess the impact of smoking cessation on outcomes following non-surgical periodontal therapy has been conducted194. This study achieved a 20% quit rate at 12 months, which confirms the effectiveness of dental healthcare professionals in providing smoking cessation counselling (Table 8)194,218,219. In this study, those individuals who successfully quit smoking had the best periodontal treatment response. The treatment response in the failed quitters (i.e. the ‘oscillators’, who initially quit smoking but then resumed) was not significantly different from that seen in those who did not manage to quit. International Dental Journal (2010) Vol. 60/No.1

The potential benefit of smoking cessation on the periodontium is likely to be mediated through a number of different pathways such as a shift towards a less pathogenic subgingival microflora, recovery of the gingival microcirculation, and improvements in aspects of the immune-inflammatory responses. In general terms, periodontal status and response to treatment in former smokers is intermediate to that seen in never smokers and current smokers, and is usually closer to that seen in never smokers. From the small number of studies to date, it appears that the periodontal status of former smokers approximates that of never smokers after around 10 years since quitting smoking. Tooth loss

Few studies have investigated the association between smoking cessation and tooth loss risk. In the study by Ahlqwist et al.89, the mean number of teeth lost in the 12 year period of follow-up was similar between never smokers and former smokers who had quit any time before baseline. In the VA DLS population, Krall et al.94 demonstrated an association between time since cessation of cigarette smoking and tooth loss risk, suggesting that it may take up to 13 years after smoking cessation for the risk of tooth loss to drop to that of never smokers. These results are consistent with the results reported by Dietrich et al.90 for the male USA health professionals, where the risk of tooth loss declined exponentially soon after smoking cessation but remained significantly elevated by about 20% even after 10+ years of smoking cessation (Figure 3). Effect of cessation on dental implants

Only one article was identified with regard to the potential effect of smoking cessation on the outcome of implant treatment220. The study reported that four out of 34 implants (12%) in subjects who refrained from smoking in conjunction with implant placement surgery showed early implant failure (implant loss or >50% bone loss) as compared to five of 13 (38%) implants in subjects who continued to smoke. Because the number of subjects involved and affected by implant loss was not given, the finding must be interpreted with great caution. Effect of cessation on other disorders

Mucosal hyperpigmentation can decrease and disappear after a successful attempt to quit smoking199. In a population of 30,118 adults from Sweden, the prevalence of oral pigmentation was found to increase during the first year of smoking. After cessation this effect resolved, and after a period of more than six years the pigmentation returned to the level found among non-smokers (≈ 3%)132,197.

21 Table 6 The effect of smoking cessation on periodontal status. Author, year

Study

Sample

Results

Bergstrom, 1991204

Cross-sectional

210 subjects (dental hygienists aged 24-60 years old). 30% current smokers, 32% former smokers, and 38% never smokers.

The mean distance from the CEJ to interproximal bone crest was significantly greater for current smokers (1.71 ± 0.08 mm) compared to never smokers (1.45 ± 0.04 mm). The corresponding measurement for former smokers was 1.55 ± 0.05 mm, intermediate to that seen in the other two groups.

Haber, 1993205

Cross-sectional

132 subjects with type 1 diabetes (28% current smokers, 13% former smokers and 59% never smokers) and 95 age- and sexstratified non-diabetic subjects (22% current smokers, 19% former smokers and 59% never smokers).

In the non-diabetic subjects aged 19-30 years, 36% of former smokers had periodontitis (defined by at least one site with probing depth ≥ 5 mm and attachment loss ≥ 2 mm) compared to 12% of never smokers and 46% of current smokers. In those aged 31-40 years, 50% of former smokers had periodontitis, compared to 33% of never smokers and 88% of current smokers. In both age groups, the difference between current and never smokers was statistically significant. A similar pattern was also seen in the diabetic patients. There were no differences between current and never smokers in the proportion of sites with plaque.

Jette, 1993109

Cross-sectional

1156 community-dwelling adults aged ≥70 years. Of the men, 18% were current smokers, and 65% were former smokers. Of the women, 8% were current smokers, and 37% were former smokers.

Duration of exposure to tobacco products (current smokers versus former smokers) was a statistically significant risk factor for tooth loss and periodontal disease, regardless of other social and behavioural factors.

Bolin, 1993206

Prospective cohort, 10 year follow-up

349 subjects examined at two intervals, 10 years apart.

Bone loss was calculated as percent loss in bone height from the first examination to the second examination. Bone loss in smokers over the 10 years was 6.0%, compared to 3.9% in never smokers. Former smokers (i.e. those who were smoking at examination one, but who had quit by examination two) demonstrated 4.4% bone loss, which was significantly less than that in the smokers.

Tomar, 2000193

Cross-sectional

12,329 dentate subjects ≥18 years old from NHANES III, of whom 27.9% were current smokers and 23.3% were former smokers.

Current smokers were 4 times more likely (odds ratio 3.97) and former smokers nearly twice as likely (odds ratio 1.68) than never smokers to have periodontitis, after adjusting for potential confounding variables. In the current smokers, there was a highly significant dose-response relationship between cigarettes smoked per day and the odds of having periodontitis. In former smokers, the odds of having periodontitis declined with the number of years since quitting, from an odds ratio of 3.22 for 0 to 2 years since quitting to an odds ratio of 1.15 (95% CI, 0.83-1.60) after ≥ 11 years since quitting. The authors concluded that 41.9% of periodontitis cases in the U.S. adult population were attributable to current cigarette smoking and 10.9% to former smoking.

Bergstrom, 2000207

Cross-sectional

257 dentally aware adults aged 20-69 years. 50 current smokers, 61 former smokers and 133 non-smokers

The mean frequency of sites with probing depths ≥ 4 mm was 6.6 in the former smokers compared with 16.8 in the current smokers and 5.2 in the never smokers (significantly greater in the current smokers compared to the other two groups). Periodontal bone height (% of root length with bone) was 82.0% in former smokers compared to 81.3% in current smokers and 84.5% in never smokers (significantly lower in the current smokers compared to the other two groups). The periodontal status of former smokers was thus intermediate to that of current and never smokers.

Table 6 continued on next page ...

Warnakulasuriya et al.: Oral health risks of tobacco use and effects of cessation

22 Table 6 continued ...... Bergstrom, 2000191

Prospective cohort, 10-year follow-up

257 dentally aware adults examined in 1982 and 1992. 10-year follow-up of data from Bergstrom 2000207.

The frequency of sites with probing depths ≥ 4 mm decreased from 11.1% to 7.8% in the former smokers, compared with a decrease from 8.7% to 6.6% in never smokers and an increase from 18.7% to 41.6% in the current smokers over the 10 year period (the frequency of such sites was significantly greater in the current smokers compared to the other two groups at the second examination). Periodontal bone height decreased from 80.7% to 79.6% in former smokers, compared with a decrease from 85.1% to 84.1% in never smokers and a decrease from 80.3% to 76.5% in current smokers. Overall, the periodontal status of former smokers was similar to that of non-smokers and remained stable (in terms of probing depths and bone height), suggesting that smoking cessation is beneficial to periodontal health.

Haffajee, 2001208

Cross-sectional

289 patients aged 20-86 years with chronic periodontitis.

Current smokers had significantly more attachment loss, missing teeth, deeper pockets and fewer sites exhibiting bleeding on probing than former or never smokers. Increasing age and smoking status were significantly related to mean attachment level and the effect of these parameters was additive.

Jansson, 200293

Prospective cohort, 20-year follow up

507 individuals who were examined in 1970 and 1990.

Interproximal bone levels were assessed from radiographs obtained in 1970 and 1990. In 1970, 50.7% were smokers and this decreased to 31.0% in 1990. Smoking was significantly correlated with alveolar bone loss over the 20 years. Individuals who quit smoking between 1970 and 1990 lost significantly less bone than those who smoked continuously.

Baljoon, 2004209

Prospective cohort, 10-year follow-up

257 dentally aware adults examined in 1982 and 1992. Same population as Bergstrom 2000207.

The prevalence of vertical bone defects in 1982 was 47% for current smokers, 49% for former smokers, and 24% for non-smokers. In 1992, the prevalence was 42%, 28%, and 19% for current smokers, former smokers, and non-smokers, respectively. Both in 1982 and 1992 the prevalence of vertical defects was significantly related to smoking status.

Paulander, 2004210

Prospective cohort, 10-year follow-up

295 subjects who were 50 years old when examined in 1988, and who were examined again in 1998.

The relative risk for a mean loss of alveolar bone height ≥ 0.5 mm was 3.69 (95% CI: 2.33-5.85) in current smokers and 0.70 (95% CI: 0.31-1.59) in former smokers

Baljoon, 2005192

Prospective cohort, 10-year follow-up

91 individuals examined in 1982 and 1992. 24 smokers, 24 former smokers, and 43 non-smokers. Sub-group of population from Bergstrom 2000207.

A statistically significant increase in the proportion of vertical defects was observed in all groups but was particularly associated with smoking. In particular, the difference between smokers and never smokers was significant, whereas former smokers did not differ from never smokers.

Torrungruang, 2005211

Cross-sectional

1960 subjects aged 50-73 years. 14.4% current smokers, 36.9% former smokers, and 48.7% never smokers.

Current smokers had more plaque, greater probing depths and more loss of attachment than former smokers and never smokers. Current smokers were 4.8 times more likely to have severe periodontitis than never smokers and former smokers were 1.8 times more likely. Quitting smoking reduced the odds of having periodontitis. For quitters who had < 15 packyears of smoking, their odds for severe periodontitis reverted to the level of non-smokers when they had quit smoking for ≥10 years. For those with ≥ 15 pack-years of smoking, the odds of having severe periodontitis did not differ from those of non-smokers when they had quit smoking for ≥ 20 years.

Thomson, 200741

Prospective cohort, 6 year follow up

810 individuals in a longstanding cohort study examined at age 26 and age 32. 31.5% were current smokers, 17.4% were former smokers, and 51.1% were never smokers.

Smokers had 5-7 times increased risk for having at least one site with ≥ 5 mm attachment loss compared to never smokers. Two-thirds of new cases of periodontitis after age 26 were attributable to smoking. There were no significant differences in periodontal health between never smokers and those who had quit smoking after age 26.

Do, 200832

Cross-sectional

3161 individuals examined as part of an Australian national oral health survey. 15% current smokers, 23% former smokers, and 62% never smokers.

The overall prevalence of periodontitis was 23%. In unadjusted analyses, former and current smokers had significantly higher prevalence of periodontitis than never smokers. Relative to non-smokers, adjusted prevalence ratios (95% CI) for periodontitis were as follows: former smokers: 1.22 (1.031.46), moderate smokers (5-15 pack years): 1.63 (1.16-2.30); and heavy smokers (> 15 pack years): 1.64 (1.27-2.12).

International Dental Journal (2010) Vol. 60/No.1

23

Table 7 The effect of smoking cessation on periodontal treatment outcomes. Author, year

Study

Sample

Results

Kaldahl, 1996212

Prospective cohort, 7-year follow-up

74 patients with periodontitis. 31 heavy smokers, 15 light smokers, 15 former smokers and 18 never smokers.

Heavy and light smokers demonstrated less favourable treatment outcomes than former smokers or never smokers.

Grossi, 1997213

Cross-sectional (3 month evaluation of treatment outcomes)

143 patients with periodontitis. 60 current smokers, 55 former smokers and 28 never smokers.

Mean probing depth reductions in former smokers were 0.49 mm (full mouth) and 1.7 mm (pockets > 5mm at baseline), almost identical to the probing depth reductions seen in never smokers (0.49 mm and 1.8 mm, respectively), and better than those observed in current smokers (0.33 mm and 1.3 mm, respectively). The probing depth reductions were significantly lower for the current smokers compared to the other two groups.

Ryder, 1999214

Prospective cohort, 9-month follow up of patients from two 9-month multicentre clinical trials that compared scaling and root planing to local delivery of doxycycline

358 patients with periodontitis. 121 current smokers, 137 former smokers, and 100 never smokers.

In the scaling and root planing group, mean probing depth reductions in former smokers (1.05 mm in pockets ≥ 5 mm and 1.58 mm in pockets ≥ 7 mm) were almost identical to those seen in current smokers (1.02 mm and 1.48 mm, respectively) and were significantly less than those observed in never smokers (1.43 mm and 2.06 mm, respectively). In the doxycycline treated group, there were no marked significant differences in clinical attachment gain or probing depth reduction among the 3 smoking groups.

Preshaw, 1999215

Prospective cohort, 6 month follow-up

41 patients with periodontitis. 15 current smokers, 14 former smokers, and 12 never smokers.

Current smokers, former smokers and never smokers did not differ significantly in plaque scores, bleeding scores, attachment levels, or radiographic bone loss. Mean probing depths were significantly greater in current smokers than in never and former smokers. Probing depths were reduced comparably in all 3 smoking subgroups following treatment, with no significant differences between groups (mean full mouth probing depth reductions were 0.55 mm in never smokers, 0.67 mm in former smokers and 0.78 in current smokers).

Meinberg, 2001216

Prospective cohort, 1-year follow-up

95 patients with treated chron- Mean baseline radiographic interproximal bone loss ic periodontitis in a 3-monthly in former smokers (4.89 mm) was intermediate to that seen in current smokers (5.75 mm) and never smokers maintenance programme. (4.64 mm). Over 1 year, bone loss occurred in 5% of the sites, but was not significantly different between the smoking sub-groups.

Hughes, 2006217

Prospective cohort, 10 weeks follow-up

79 patients with generalised aggressive periodontitis. 20 were current smokers, 19 were former smokers and 40 were never smokers.

Treatment outcomes were poorer in smokers (mean probing depth reduction 1.75 ± 0.56 mm) than in the former smokers and never smokers combined (mean probing depth reduction 2.23 ± 0.87 mm). There were no significant differences in treatment responses between former smokers (mean probing depth reduction 2.1 ± 0.84 mm) and never smokers (mean probing depth reduction 2.3 ± 0.89 mm).

Warnakulasuriya et al.: Oral health risks of tobacco use and effects of cessation

24 Table 8 Interventional studies of the effect of smoking cessation on the periodontium. Author, year

Study

Sample

Results

Nair, 2003218

Interventional study, short term (4-6 weeks) investigation of effect of smoking cessation on gingival bleeding

27 smokers participating in a quit smoking programme who had no signs of advanced periodontitis.

Data for the 27 quitters revealed a statistically significant increase in the proportion of sites that bled on (constant force) probing from 15.7% to 31.9% as a result of quitting. At the same time, a significant decrease in the proportion of sites with plaque occurred, from 38.9% to 28.1%.

Morozumi, 2004219

Interventional study, short term (8 weeks) investigation of effect of smoking cessation on gingival vasculature

16 male, periodontally healthy 11 of the 16 subjects successfully quit smoking for smokers. 8 weeks (and the other 5 subjects dropped out of the study). Gingival blood flow (determined by laser Doppler flow) and gingival crevicular fluid volume increased significantly following smoking cessation in the 11 quitters.

Preshaw, 2005194

Interventional study, 12 month 49 smokers with chronic follow-up periodontitis who wished to quit smoking.

After 12 months, of patients with complete data, 10 had continuously quit (20% quit rate), 10 continued smoking and six were ‘oscillators’ (patients who quit and then resumed smoking). Analysis of probing depth reductions between baseline and month 12, and comparing quitters with the other two groups combined, demonstrated a significant benefit of quitting on mean probing depth reductions (additional mean reduction of 0.32 mm in the quitters). Quitters were significantly more likely to demonstrate probing depth reductions ≥ 2 and ≥ 3 mm than non-quitters and oscillators. Thus, quitters had increased odds (relative risk = 1.67) of demonstrating probing depth reductions ≥ 2 mm compared with the other two groups, and more than twice the odds of demonstrating probing depth reductions ≥ 3 mm (relative risk = 2.36).

Time-dependant association of smoking cessation with risk of tooth loss (Reference: never smokers)

Hazard-Ratio (95% CI)

4

3

2

1

0